Skills
skills/jeremylongshore/claude-code-plugins-plus-skills/generating-infrastructure-as-code/Gen Agent Trust Hub

generating-infrastructure-as-code

Warn

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The Python script scripts/iac_generate.py includes a generate_script function that assembles shell scripts from template content provided via arguments and explicitly sets executable permissions (chmod 0o755). This allows the agent to generate and potentially execute arbitrary code on the local system.
  • [PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface by processing natural language infrastructure requirements while possessing broad access to cloud management CLIs.
  • Ingestion points: Architectural specifications provided in user prompts.
  • Boundary markers: None specified in SKILL.md to isolate user input.
  • Capability inventory: Access to Bash for terraform, aws, gcloud, and az tools.
  • Sanitization: No input validation or escaping is implemented before processing requirements.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 28, 2026, 07:06 PM